A major role for zygotic hunchback in patterning the Nasonia embryo

The patterning of insect embryos is controlled by a spectrum of well-conserved to rapidly evolving genes. The molecular genetics of axis formation has been examined in a variety of insect embryos(Tautz and Sommer, 1995; Dearden and Akam, 1999; Lall and Patel, 2001; Lynch and Desplan, 2003b) and provides a framework for exploring the fundamental principles of regulatory gene evolution.

Anteroposterior axis formation is best understood in Drosophila,where early embryogenesis takes place extremely rapidly and depends heavily on maternal input (St. Johnston and Nüsslein-Volhard, 1992; Rivera-Pomar and Jäckle,1996). The bicoid homeodomain morphogen is provided maternally as mRNA localized to the anterior of the oocyte(Berleth et al., 1988). Bicoid synergizes with Hunchback, a zinc-finger protein, in controlling anterior development (Simpson-Brose et al.,1994). hunchback, a gap gene, transcriptionally controls other gap genes, as well as pair-rule and homeotic genes(Pankratz and Jäckle,1993; Simpson-Brose et al.,1994; Tautz and Sommer,1995; Casares and Sánchez-Herrero, 1995; Fujioka et al., 1999; Shimell et al., 2000; Wu et al., 2001; Clyde et al., 2003). hunchback, in contrast to bicoid, is provided maternally as unlocalized mRNA, and is expressed zygotically under the control of bicoid and other transcriptional regulators(Bender et al., 1988; Schröder et al., 1988; Tautz, 1988; Margolis et al., 1995). Although not essential, maternal hunchback does control some head-determining functions in wild-type Drosophila, as embryos lacking all maternal and zygotic products have a larger anterior gap than those lacking only zygotic hunchback(Lehmann and Nüsslein-Volhard,1987). Maternal hunchback must be translationally repressed by nanos for normal posterior development(Hülskamp et al., 1989; Irish et al., 1989; Struhl, 1989).

To better understand the evolution of anteroposterior patterning, we have taken advantage of the haplo-diploid genetic system of the wasp Nasonia vitripennis to screen for mutations affecting cuticular morphology. In haplo-diploids, fertilized eggs develop as diploid females while unfertilized eggs develop as haploid males, facilitating a screen of the genome for recessive zygotic mutations (Pultz and Leaf, 2003). We identified about one-fourth to one-third of the genes required to pattern the Nasonia embryo, including representatives of gap, pair-rule and Polycomb-group genes with varying degrees of functional similarity to known Drosophila genes(Pultz et al., 2000). Three zygotic mutations caused extensive disruptions of early patterning, more severe than the defects caused by any known zygotic mutation in Drosophila. One of the Nasonia mutations, originally named headless, deletes all of the head except the most anterior labral segment, as well as thoracic and posterior abdominal segments(Fig. 1A)(Pultz et al., 1999,). The headless mutant phenotype suggested a similarity to Drosophila hunchback. Zygotic loss of hunchback in Drosophila causes a gap deletion from the posterior labial segment– the posterior border of the head – through the thoracic segments, and also affects posterior abdominal segments(Fig. 1A)(Bender et al., 1987; Lehmann and Nüsslein-Volhard,1987). We hypothesized that the headless mutant phenotype is caused by a mutation in Nasonia hunchback, and that zygotic hunchback plays a more extensive role in embryonic patterning in Nasonia than in Drosophila(Pultz et al., 1999).

Non-dipteran insects must initiate embryonic patterning using different methods from those of Drosophila. Although bicoid controls the development of head, thorax and anterior abdomen in Drosophila,the bicoid gene has apparently arisen only relatively recently,within the higher Diptera (Stauber et al.,1999; Brown et al.,2001; Stauber et al.,2002). bicoid encodes a homeodomain protein with a key lysine at position fifty (K50) of the homeodomain, and is hypothesized to have usurped functions originally controlled by orthodenticle, which also encodes a K50 homeodomain protein; in addition, hunchback is hypothesized to have played a more extensive role in patterning the anterior of more ancestral insects (Wimmer et al.,2000; Lynch and Desplan,2003a). Parental RNA interference experiments in the beetle Tribolium have indicated that the orthodenticle gene plays a major role in patterning the anterior of this non-Dipteran insect; when both Tribolium orthodenticle and Tribolium hunchback(Wolff et al., 1995) are knocked down, very little remains of the segmental patterning in the Tribolium embryo (Schröder,2003). The potential role of hunchback as an ancestral morphogen has also been tested by manipulating hunchback expression in Drosophila, where increased levels of bicoid-independent hunchback have been shown to be capable of patterning the abdomen and even the thorax in the absence of bicoid(Hülskamp et al., 1990; Struhl et al., 1992; Schulz and Tautz, 1994; Wimmer et al., 2000).

Studies of hunchback in representatives of more ancestral insects have provided an intriguing perspective on the evolution of this key regulatory gene. In the milkweed bug Oncopeltus, hunchback mRNA is expressed both maternally and zygotically, and embryos with knocked-down hunchback function exhibit transformations of gnathal and thoracic segments to an abdominal identity, as well as impaired germ-band development(Liu and Kaufman, 2004). By contrast, in the more ancestral grasshopper, Schistocerca, hunchbackis provided maternally to the embryo as protein, rather than as mRNA, through release from the posteriorly located oocyte nucleus, suggesting that its function may be to distinguish embryonic from extra-embryonic cells in that short-germ embryo (Patel et al.,2001). A later graded expression of SchistocercaHunchback is provided zygotically, consistent with a concentration-dependent role in axial patterning. These results indicate that the ancestral hunchback axis-determining function in insects is likely to be zygotic, and that the expression and function of maternal hunchbackhas significantly changed during insect evolution.

Here, we focus on the role of hunchback in the hymenopteran Nasonia vitripennis. Nasonia has long-germ embryos, with a syncytial mode of early development that is morphologically similar to Drosophila embryogenesis – although unlike Drosophila,Nasonia does not have the highly derived condition of extremely rapid early development. In fact, approximately three-fold more time is allocated to early development (prior to gastrulation) in Nasonia than in Drosophila (Fig. 1B)(Bull, 1982; Campos Ortega and Hartenstein,1985) – allowing more time for the zygotic genome to control early development. The Hymenoptera have evolved diverse embryonic developmental strategies. These include embryos with derived holoblastic cleavage (Grbic and Strand,1998), as well as several independently evolved cases of polyembryony, in which a single fertilized egg develops into hundreds or thousands of clonal progeny (Strand and Grbic, 1997; Grbic,2000). Syncytical long-germ development is considered to be ancestral in the Hymenoptera (Strand and Grbic, 1997), so Nasonia can be considered to be a representative of the ancestral mode of development within this clade.

We show that the severe Nasonia headless zygotic-mutant phenotype is caused by a mutation in Nasonia hunchback, and we describe the expression of Nasonia hunchback mRNA and protein. We also compare molecular mutant phenotypes of Nasonia embryos lacking zygotic hunchback to those of Drosophila embryos lacking both maternal and zygotic hunchback. We propose that the divergent mutant phenotypes for the same gene in two different species may, in large part, be due to changes in the functionally overlapping genetic regulatory network.

The highly conserved middle zinc-finger region of Nasonia hunchback was cloned from Nasonia vitripennis (Nv) and from Nasonia giraulti (Ng) using degenerate forward(5′-CGCGAATTCAARCAYCAYCTNGARTAYCA-3′) and reverse primers(5′-ATATGCGACRTGRCARTAYTTNGTNGCRTA-3′) with the following PCR cycling conditions: 94°C for 30 seconds, 60°C for 60 seconds and 72°C for 120 seconds, for 32 cycles.

To identify a Ng-specific hunchback single nucleotide polymorphism, a Ng-specific forward primer(5′-CCATCTGCGCAAGCA-3′), and a species non-specific reverse primer(5′-GCAGTCGCAGCACCT-3′) were used to amplify a Ng hunchback fragment with the following PCR cycling conditions: 95°C for 30 seconds, 58°C for 60 seconds, 72°C for 60 seconds for 33 cycles.

To determine linkage, Nv headless-bearing females were crossed to Ng males cured of Wolbachia with antibiotics, kindly provided by Jack Werren (University of Rochester, NY, USA). The Nv headless/Ng headless⁺ F1 hybrids were sorted from their Nv headless⁺/Ng headless⁺control sisters by assaying their embryos. Experimental and control females were set unmated, then DNA was prepared from single surviving adult F2 males(Gloor et al., 1993). DNA that failed to amplify with the Ng-specific primer was shown to support amplification with species non-specific primers.

The deletion in Nasonia hunchback was characterized by isolating DNA from 30-50 selected headless mutant embryos(Gloor et al., 1993). PCR amplification with primers at the 5′ and 3′ ends of the coding region generated a product approximately 1.5 kb shorter than the wild-type product, indicating a deletion. The mutant product was cloned and sequenced. The precise size of the deletion was 1497 bp, consistent with the PCR analysis. Identical sequences across the breakpoint were obtained from two independently amplified reactions.

When Nasonia embryos are collected from virgin females, all embryos are precisely staged – there are no older embryos from previously fertlized eggs as in Drosophila. Embryos after gastrulation were fixed as described in Pultz et al.(Pultz et al., 1999). Most of the blastoderm embryos were shaken in heptane for 2 minutes, then an equal volume of methanol was added and they were shaken for an additional 2-3 minutes at room temperature. Later, we found that sufficiently dry blastoderm embryos can also be fixed in 1:1 heptane: 4% formaldehyde in 1×PBS,improving morphology. Very early embryos (0-3 hours old) cannot be effectively devitellinated with methanol. These were fixed for 1 hour in heptane pre-saturated with 37% formaldehyde, then hand-peeled on double-stick tape in 1×PBS. Older hand-peeled embryos with a known expression pattern were included as a positive control. All embryos were males, collected from virgin mothers. To avoid cross reactivity of the anti-Nasonia hunchbackantibody with endosymbiotic bacteria, we used wild-type Nasonia vitripennis cured of Wolbachia (a gift from Jack Werren), and we cured the hunchback^hl stock of Wolbachia by treating the mothers for two generations with rifampicin. Ovaries were dissected from mothers, fixed for 10 minutes in 8% formaldehyde, dehydrated and stored in methanol or ethanol until used for antibody staining or in situ hybridization, respectively.

Nasonia hunchback mRNA was visualized using an anti-sense RNA probe, as described previously (Jiang et al., 1991). The probe, about 1100 bp in length, extended from exon 2 through the central zinc-finger region (see Fig. 3). As a negative control,a probe was prepared from the opposite strand of the same fragment, and was applied to samples of all ages of embryos and tissues analyzed. No staining was observed with the negative controls.

A 125-amino-acid region beginning at amino acid 79 and terminating before the NF1 zinc finger was PCR-amplified using forward(5′-GTTGTTGAATTCGCTGGGATAAAATCGTA-3′) and reverse(5′-GTTGATAAGCTTGGGCAGCTCGAATCC-3′) primers, then cloned into the EcoR1 and HindIII sites of pGEX-KG, producing a GST-hunchback fusion protein. The fusion protein was isolated as described by Leaf and Blum (Leaf and Blum,1998) and injected into rabbits for the production of polyclonal antiserum.

The anti-Nasonia Hunchback antibodies were used at a dilution of 1:1000 to stain Nasonia embryos. All staining patterns observed in wild-type embryos – of cellular blastoderm age and older – were verified to be absent in hunchback^hl mutant embryos. The FP6.87 monoclonal antibody (Kelsh et al.,1994), which recognizes conserved epitopes on both Ultrabithorax(Ubx) and Abdominal-A (Abd-A) proteins was used at a dilution of 1:7 to stain Nasonia and Drosophila embryos. The anti-Drosophila hunchback guinea pig polyclonal antibody(Kosman et al., 1998) was used at a dilution of 1:400. All antibodies were visualized with horseradish peroxidase-labeled secondary antibodies and diaminobenzidine substrate, as described by Pultz et al. (Pultz et al.,1999).

To analyze maternal Hunchback expression, embryos were collected from parents heterozygous for Df (3R) p²⁵, which deletes the 5′ end of the hunchback transcription unit and does not produce hunchback mRNA (Bender et al.,1988). To analyze Ubx-Abd-A expression in embryos lacking both maternal and zygotic hunchback, we used a hb^FBFRT strain kindly provided by Ernst Wimmer (Georg-August-University Göttingen, Germany), collecting the embryos from FLP-bearing hb^FB FRT/ovo^D mothers that had been heat shocked as larvae to induce clones homozygous for the null hunchbackmutation in their ovaries (Dang and Perrimon, 1992). These females were crossed to hb^14F/TM3 males, such that approximately 50% of the offspring lacked both maternal and zygotic hunchback, whereas the other 50% lacked only maternal hunchback. Because maternal hunchback is not needed by Drosophila embryos in the presence of zygotic hunchback, half of the embryos showed a wild-type pattern of Hox gene expression. The other half, lacking both maternal and zygotic hunchback, were severely defective. The mutant phenotypes were confirmed using cuticle preparations.

We began by cloning a small genomic fragment of the Nasonia vitripennis (Nv) hunchback gene containing the four middle zinc fingers, which are highly conserved in insects(Sommer et al., 1992, Patel et al., 2001). Next, to investigate whether the Nv-hunchback sequence maps to the headless (hl) mutation, we took advantage of the fact that N. vitripennis can be crossed to a sibling species, N. giraulti (Ng). Fertile F1 hybrids can be generated by removing Wolbachia, an endosymbiont responsible for cytoplasmic incompatibility (Breeuwer and Werren,1990), which generates asynchronous cell cycles of the male and female pronuclei (Tram and Sullivan,2002). The estimated divergence time of approximately 200,000 years (Campbell et al., 1993)between the sibling species enhances the likelihood that a single nucleotide polymorphism (SNP) for mapping can be identified in a short and highly conserved sequence. After identifying such an SNP in Nasonia hunchback, we used an Ng-specific primer(Fig. 2A,B) to analyze the surviving adult male progeny from virgin Nv hl/Ng hl⁺experimental mothers and Nv hl⁺/Ng hl⁺ control mothers (Fig. 2C, Materials and methods). We found that 105/105 surviving sons of the experimental mothers were hemizygous for the Ng-hunchback⁺ allele, as would be expected if the headless mutant phenotype were due to a lesion in Nv hunchback. (In the sons of the control mothers, Nv and Ng alleles segregated approximately equally: 9 Ng to 6 Nv.) These results led us to examine the DNA of headlessmutant embryos for a lesion in the Nasonia hunchback gene.

Nasonia Hunchback shares with other hunchback proteins a set of four-conserved zinc fingers in the middle of the protein and two zinc fingers at the C terminus (Fig. 3A). In addition, Nasonia Hunchback has an additional N-terminal zinc finger, not found in Drosophila or TriboliumHunchback, which is similar to the Nf-1 zinc finger of Schistocercaand Oncopeltus Hunchback (D.S.L. and M.A.P., unpublished). As in Drosophila and Tribolium, hunchback appears to be transcribed in Nasonia from more than one promoter(Fig. 3A; D.S.L. and M.A.P.,unpublished).

To identify the headless mutation, genomic DNA from headless mutant embryos was amplified and sequenced, revealing a deletion of 1497 bp after the first 40 amino acids of the predicted protein-coding sequence (assuming that translation starts in exon 2). As shown in Fig. 3B, this deletion disrupts the reading frame for the protein-coding sequence. This most likely defines a null allele for the Nasonia hunchback gene (see Discussion), consistent with our hypothesis that the very severe headless mutant phenotype is caused by a loss of zygotic Nasonia hunchback. Consequently, we re-designated Nasonia headless(hl) as Nasonia hunchback^hl.

The translational regulation of Drosophila hunchback is mediated by the binding of Pumilio to Nanos Response Elements (NREs) within the 3′ untranslated region (UTR), and the subsequent recruitment of Nanos and Brain Tumor to form a quarternary complex(Murata and Wharton, 1995; Sonada and Wharton, 2001; Wang et al., 2002). Figure 3C shows candidate NREs from Nasonia hunchback, which are similar to the conserved Box A and Box B of the Drosophila hunchback NREs. The canonical hunchback NREs have a characteristic spacing of three to four bases between Box A and Box B. By contrast, the candidate NREs of Nasonia hunchback have 12-16 bases separating Box A and Box B, reminiscent of the structure of a candidate NRE found in the 3′ UTR of Drosophilacyclin B1 mRNA (Wang et al.,2002). In the germline, Pumilio and Nanos translationally repress Drosophila cyclin B1 expression(Nakahata et al., 2001). The presence of candidate NREs is of interest in light of the difference between the mRNA and protein expression of Nasonia hunchback described below.

To determine whether Nasonia hunchback is expressed maternally,and how embryonic expression compares with that of other insects, we examined Nasonia hunchback mRNA expression, and we raised an antibody against part of Nasonia Hunchback (HB-GST, Fig. 3). We found that hunchback mRNA is supplied to the embryo maternally, from high-level expression in the oocyte (Fig. 4), and as an mRNA that is dispersed throughout the egg (data not shown) and not yet translated, as no protein could be detected in 0-1 hour embryos (data not shown).

The timing of key morphological events during embryogenesis at 28°C is summarized in Fig. 5A. Just before pole cell formation, Nasonia Hunchback is expressed ubiquitously (data not shown). An anterior to posterior gradient of Nasonia Hunchback begins to form soon after the nuclei migrate to the surface and begin dividing at the surface of the embryo(Fig. 5C). The protein is localized to nuclei. During the next two hours (at 28°C), until the beginning of cellularization, the embryos continuously express an anterior domain of Nasonia Hunchback, with a sharpening border(Fig. 5E). However, throughout this period of graded Nasonia Hunchback expression, we did not observe a parallel gradient of Nasonia hunchback mRNA expression. Rather, the embryos express a low ubiquitous level of hunchback mRNA,superimposed with a small anterior and a larger posterior domain of expression during the cell cycles just after pole cell formation(Fig. 5B). This is followed by restriction of the posterior domain to the posterior, with incipient expression at the center of the embryo(Fig. 5D). The difference between the lack of Nasonia Hunchback at the posterior and the continuous presence of the mRNA at the posterior, throughout this two-hour period, indicates that Nasonia hunchback must be translationally controlled.

During the later stages of blastoderm development, the expression of Nasonia Hunchback appears to follow the expression of the mRNA, first expressed as anterior and posterior domains(Fig. 5F,G), then retracting from the anterior and anterodorsal region, and from the posterior. The anterior Hunchback domain diminishes in intensity in both the mRNA and protein expression prior to the onset of gastrulation, whereas the posterior stripe is still strongly expressed (Fig. 5H,I).

Just prior to gastrulation, a narrow stripe of Nasonia hunchbackmRNA expression appears on the dorsal side of the embryo(Fig. 6A). Upon germ-band extension, this dorsal expression domain appears to be associated with serosa development (Fig. 6B-D). Although in many primitive insects the serosa develops from the anterior, in Nasonia the serosa, an extra-embryonic membrane, begins to develop in the dorsal region of the embryo, then expands anteriorly and ventrally to eventually envelop the entire embryo (Bull,1982). Finally, Nasonia Hunchback is also expressed in a patterned subset of cells in the central nervous system, most strongly during the period of head involution (Fig. 6E,F).

We investigated whether Nasonia hunchback function is needed for serosa formation, by comparing living wild-type and hunchback^hl embryos (data not shown). We found that the serosa still forms apparently normally in the mutant embryos, indicating that although zygotic Nasonia Hunchback is expressed relatively early and strongly in this tissue, it is not necessary for its morphological determination.

To investigate whether maternal hunchback mRNA contributes to the anterior expression of Nasonia Hunchback, we determined whether Hunchback expression could be detected in hunchback^hlmutant embryos, which lack zygotic Hunchback. Because the anti-Nasonia Hunchback antibody was generated against a region of the protein that was completely deleted in hunchback^hl mutant embryos (Fig. 3), any protein detected with this antibody in the mutant embryos must be maternally derived. Male embryos were collected from hunchback^hl/+ virgins, so one-half of the embryos should express no zygotic Hunchback. In a collection of such embryos, aged from about cycle 10 to cellularization (4-6 hours AEL at 28°C; Fig. 5A), all embryos(55/55) expressed Hunchback in an anterior domain (see Fig. 5C,E). The youngest embryos in the collection all appeared to have a similarly strong Hunchback expression, whereas three of the oldest embryos in this collection,approaching the beginning of cellularization, had barely detectable levels of the Hunchback gradient. A control for this experiment was a collection of older embryos from the same mothers, in which only half of the embryos expressed Hunchback, as expected during the purely zygotic phase of expression(see below). These results indicate that maternally derived Nasonia hunchback mRNA contributes to a gradient of Hunchback in Nasoniaembryos prior to cellularization.

Why does a lack of zygotic hunchback result in more severe consequences in Nasonia than in Drosophila, despite graded maternal Hunchback expression in both species? We hypothesized that because of the longer period of early development in Nasonia(Fig. 1) maternal Hunchback does not overlap temporally with zygotic Hunchback to the same extent that it does in Drosophila. To test this hypothesis, we examined Hunchback in Nasonia hunchback^hl mutant embryos, and compared them with Drosophila embryos lacking zygotic Hunchback, during the period when maternal Hunchback is decaying. Specifically, we examined whether residual maternal Hunchback is detected near the onset of cellularization, when both Nasonia and Drosophila embryos begin to express Hunchback zygotically in a posterior cap (in addition to the anterior domain).

In a tightly staged collection of male Nasonia embryos from hunchback^hl/+ virgins, we observed 34 embryos expressing Hunchback in the anterior and incipient posterior caps(Fig. 7A), while 31 sibling embryos had no detectable Hunchback expression(Fig. 7B). In a control experiment, all of 50 Nasonia wild-type embryos of a similar age clearly showed the zygotic Hunchback expression pattern. These results show that in Nasonia embryos, maternal Hunchback does not persist into the period of posterior cap expression, but our characterization of maternal Hunchback in earlier embryos (see the previous section above) indicates that it is weakly expressed just prior to that time.

To examine maternal Hunchback in Drosophila, we made synchronous collections of embryos from hunchback^–/+ parents(see Materials and methods) and from wild-type parents, such that the youngest embryos were just beginning to express the zygotic posterior Hunchback cap(Fig. 7C). The progeny of the wild-type parents all expressed Hunchback in a strong anterior domain, as well as in the posterior cap. However, in 25% of the progeny of heterozygous parents (57/228), we found either weak staining only in anterior nuclei– presumably from residual maternal expression – or no detectable staining. Specifically, we observed 27 progeny of the heterozygous parents with only weak anterior staining (maternal expression only; Fig. 7D) and 30 with no staining. The 90 youngest siblings with strong anterior staining (mostly zygotic expression) showed incipient posterior-cap staining. (The remaining 81 siblings with zygotic expression were older, exhibiting either strong posterior cap staining or resolution of the posterior cap into a posterior stripe.) Because 27 is close to one-fourth of 117 (27 maternal plus 90 youngest zygotic), these results indicate that in Drosophila,maternal Hunchback perdures into the period of incipient zygotic posterior cap expression, during early cellularization. This is in contrast to Nasonia maternal Hunchback, which appears to decay before the equivalent stage of posterior cap expression during early cellularization. This timing difference may contribute to Nasonia's stronger dependence on zygotic hunchback. However, the experiments described below indicate that maternal hunchback cannot fully account for the difference in functions covered by zygotic hunchback in Nasonia and Drosophila.

To better understand the greater essential zygotic role of hunchback in Nasonia than in Drosophila, we examined the effects of hunchback mutant genotypes on Hox gene expression. In previous work (Pultz et al., 1999), we had compared Nasonia hunchback^hl to Drosophila embryos lacking zygotic hunchback for their effects on trunk Hox gene expression, Ultrabithorax (Ubx) and abdominal-A(abd-A), and demonstrated that the ectopic expression of Ubx-Abd-A extends more anteriorly in the Nasonia headless mutant embryos. Here,we extend this comparison to Drosophila embryos lacking both maternal and zygotic hunchback. For these analyses, as previously, we used the phylogenetically cross-reactive monoclonal antibody generated by Kelsh et al.(Kelsh et al., 1994), which recognizes epitopes on both Ubx and Abd-A. Expression of Ubx-Abd-A in a wild-type Drosophila embryo is shown in Fig. 8A; expression in a Drosophila embryo lacking zygotic hunchback is shown in Fig. 8B. As reported previously(White and Lehmann, 1986; Pultz et al., 1999), the trunk homeotic genes are derepressed both anteriorly and posteriorly when zygotic Drosophila hunchback function is eliminated. This derepression does not extend anteriorly into the maxillary segment. By contrast, Fig. 8C shows Ubx-Abd-A expression in Drosophila embryos lacking both maternal and zygotic hunchback, generated using germ line clones(Dang and Perrimon, 1992). In these embryos, no maxillary lobe develops, and the Hox gene expression extends slightly further anteriorly than with loss of only zygotic hunchbackfunction, but much of the head is still clear of the trunk Hox gene expression. In wild-type Nasonia embryos, the trunk Hox gene expression (Fig. 8D) is very similar to that of wild-type Drosophila embryos. In hunchback^hl mutant embryos, as in Drosophilaembryos lacking hunchback zygotic function, the Hox genes are derepressed anteriorly as well as posteriorly(Fig. 8E,F). However, the ectopic expression of trunk Hox genes in Nasonia hunchback^hl mutant embryos extends much further anteriorly than in Drosophila embryos lacking zygotic hunchback, and even appears to extend further anteriorly than in Drosophila embryos lacking both maternal and zygotic hunchback. This indicates that zygotic hunchback in Nasonia controls more functions than all hunchback, both maternal and zygotic, in Drosophila.

We show here that the Nasonia headless mutant phenotype, which we described in Pultz et al. (Pultz et al.,1999), is caused by a 1.5 kb deletion in the Nasonia hunchback gene. This deletion begins after 40 amino acids of the predicted reading frame and introduces a frameshift mutation, disrupting the remaining reading frame such that the mutant protein lacks all zinc fingers(Fig. 3). Therefore, we have renamed headless (hl) as hunchback^hl. Hülskamp et al. (Hülskamp et al., 1994) describe several amorphic (functionally null) alleles of Drosophila hunchback including hb^FB, a 10 bp deletion that introduces a frameshift mutation after the first 150 amino acids, and hb^14F, which introduces a stop codon at amino acid 236. As in Nasonia Hunchback^hl, DrosophilaHunchback^FB and Hunchback^14F lack all zinc fingers. These comparisons indicate that Nasonia hunchback^hl can be considered to be a null allele.

The hunchback^hl mutant phenotype (hemizygous progeny of heterozygous mothers) reveals that zygotic hunchback is essential in Nasonia for development of almost the entire head, as well as the thorax and the posterior abdomen. In Drosophila, when only zygotic hunchback is removed (in homozygous progeny of heterozygous parents),the posterior labial segment and thorax – plus a small posterior abdominal region – are deleted, but underlying maternal hunchback still patterns part of the head. When both maternal and zygotic hunchback are removed, the anterior defects expand further into the head (Bender et al.,1987; Lehmann and Nüsslein-Volhard, 1987).

The above comparison of mutant phenotypes, the observation that more time is allocated to early development in Nasonia, and evidence from Schistocerca that the hunchback axial patterning function may originally have been zygotic (Patel et al., 2001), together suggested that hunchback might be expressed only zygotically in Nasonia. However, we have found hunchback mRNA in ovaries and in very early embryos, indicating that hunchback is transcribed maternally in Nasonia, as in Drosophila and Tribolium(Wolff et al., 1995), although it is not translated maternally as in Schistocerca(Patel et al., 2001). Moreover, by examining Hunchback expression in hunchback^hlmutant embryos, we found that maternal hunchback mRNA appears to be solely or primarily responsible for directing the synthesis of the early anterior Hunchback domain during the first cell cycles after pole cell formation in Nasonia embryos.

The finding that Nasonia hunchback is expressed maternally –even though zygotic hunchback controls more extensive patterning in Nasonia than in Drosophila – raises the question:what, if any, is the function of maternal hunchback in Nasonia? One possibility is that maternal hunchback is necessary as a positive regulator of zygotic hunchback; for example,in Drosophila, the parasegment 4 expression of hunchback is under positive regulation by hunchback gene products(Hülskamp et al., 1994; Margolis et al., 1995). If autoregulation were the sole role of maternal hunchback, then eliminating both maternal and zygotic hunchback would produce the same defects as eliminating only the zygotic gene products. Attempts to eliminate both maternal and zygotic hunchback function in Nasonia with parental RNA interference have only rarely yielded embryos with a phenotype as strong as that of hunchback^hlmutant embryos (J.L. and C.D., unpublished). This suggests that the maternal gene products may not control additional anteroposterior patterning functions.

In contrast to the largely conserved expression of Nasonia hunchback protein (see below), the expression of Nasonia hunchback mRNA differs from that of Drosophila. In Drosophila, hunchback maternal mRNA at the posterior of the embryo degrades as the mRNA is being translationally controlled by Nanos, generating a gradient in both the mRNA and the protein expression(Bender et al., 1988; Schröder et al., 1988; Tautz, 1988). Similar posterior degradation of the maternal mRNA has also been observed in the housefly Musca domestica (Sommer and Tautz, 1991). By contrast, throughout the two-hour period when Nasonia embryos are expressing an anterior Hunchback domain (from just after the nuclei arrive at the surface until the onset of cellularization), there is a substantial domain of hunchback mRNA at the posterior end of the embryo. The absence of Hunchback at the posterior of the embryos indicates that Nasonia hunchback is under translational control, presumably by Nanos. Schistocerca hunchback also appears to be translationally controlled at the posterior of the embryo(Lall et al., 2003). Nasonia hunchback does not have a canonical NRE such as is found in Schistocerca, Locusta, Tribolium and Drosophila hunchbackmRNAs; however, Nasonia hunchback does have candidate NREs that are similar in structure to the Drosophila melanogaster cyclin B1 NRE,which is translationally regulated by Pumilio and Nanos in the germline(Nakahata et al., 2001).

Because Nasonia and Drosophila differ in the extent of essential zygotic hunchback function and in the timeline for early development, we compared the overall timing, as well as pattern of Hunchback expression in the wild-type embryos. Our comparative observations of wild-type Hunchback expression in Drosophila (not shown, see Materials and methods) and Nasonia indicate that the dynamics of expression largely correlate with the same morphological markers during blastoderm development,rather than with absolute developmental time. In both Nasonia and Drosophila, just before pole cell formation, Hunchback is expressed ubiquitously, then this expression is replaced by an anteroposterior gradient with a progressively sharpening border as the cells begin dividing at the surface of the embryo. Very early in the process of cellularization, in both Nasonia and Drosophila embryos, a posterior cap of Hunchback begins to be expressed in addition to the anterior domain of expression. In both organisms, during the cellular blastoderm period, the posterior cap resolves into a posterior stripe and expression retracts from the anterior. However, the expression patterns of Hunchback in Nasonia and Drosophila are not entirely identical. The anterior domain of Hunchback persists into the early stages of germ band extension in Drosophila but not in Nasonia, and at gastrulation, a dorsal stripe is expressed in Nasonia that is not present in Drosophila.

Dorsal Nasonia Hunchback expression appears to be associated with development of the serosa. Expression in extraembryonic membranes is an aspect of Hunchback expression that has been described for other insects, including Schistocerca and Drosophila(Patel et al., 2001), as well as the mosquito Anopheles gambiae(Goltsev et al., 2004). We find that despite zygotic expression in the serosa, Nasonia zygotic hunchback function is not necessary for serosa formation. The dorsal expression domain suggests that zygotic hunchback might be positively regulated in Nasonia by the zerknüllt (zen)homeobox gene. In Drosophila, zen is dorsally expressed during blastoderm development and is required for development of the amnioserosa(Wakimoto et al., 1984; Doyle et al., 1986). zen is also expressed in the anterior and dorsal regions that give rise to extraembryonic membranes in other insects(Falciani et al., 1996; Stauber et al., 2002); in Tribolium, anterior expression of zen1 has also been shown to specify serosa cell fates, differentiating them from those of the more posterior germ rudiment (van der Zee et al., 2005). Nasonia Hunchback is also strongly expressed in the nervous system, approximately during the period of head involution, in a pattern that appears to be similar to that observed in the nervous system of other insects (Woff et al., 1995; Rohr et al., 1999; Patel et al.,2001).

Because maternal hunchback is partially redundant with zygotic hunchback in Drosophila(Lehmann and Nüsslein-Volhard,1987), we also examined the timing of the maternal component relative to the zygotic component in Nasonia and Drosophila. We found that maternal Hunchback expression in Drosophila appears to overlap with a slightly later zygotic phase of expression than in Nasonia. This timing difference may contribute to Nasonia's greater reliance on zygotic hunchback. However, the strikingly different essential roles of hunchback in Nasonia and Drosophila call for further explanation.

We have considered several possible explanations to account for the observation that the zygotic hunchback loss-of-function phenotype is more severe in Nasonia than in Drosophila. As discussed above, we first hypothesized that Nasonia lacks maternally provided hunchback function, but this explanation was ruled out, as Nasonia does have strong maternal Hunchback expression. Second, we found that more limited perdurance of maternal Hunchback during the blastoderm stage of Nasonia may contribute to the differential function. Third,we consider here that Nasonia Hunchback might also regulate more downstream genes, either by DNA-binding or protein-protein interactions, than Drosophila Hunchback. In this regard, it is notable that Nasonia Hunchback has an N-terminal zinc finger (NF-1) that is lacking in Drosophila. However, the function of NF-1 is not understood, and N-terminal zinc fingers of Hunchback have been independently discarded in number of insect taxa including Hymenoptera (Apis) and Orthoptera(Cricket). Finally, Nasonia and Drosophila may differ in the degree to which other genes are redundant or synergistic with Hunchback function.

Our analysis of Hox gene expression in Drosophila embryos indicated that even when both maternal and zygotic hunchback products are removed, the defects are not as extensive as the zygotic defects of Nasonia hunchback^hl. Consistently, cuticular analyses of Drosophila embryos lacking both maternal and zygotic hunchback show that the deleted region extends forward only through the maxillary segment (E. Wimmer, personal communication); however, all gnathal plus at least two pregnathal segments are deleted in Nasonia hunchback^hl (Pultz et al.,1999). This raises the question of whether the absence of a bicoid gene in Nasonia could potentially be responsible for the extent of the defects observed with a loss of zygotic Nasonia hunchback.

When the dose of maternal bicoid was reduced by half in Drosophila embryos that also lacked all maternal and zygotic hunchback (E. Wimmer and C.D., unpublished), the array of head segments deleted (all except the labrum) was identical to the region deleted in Nasonia hunchback^hl mutant embryos. Importantly, these`headless' Drosophila mutant embryos can be rescued by a single zygotic hunchback⁺ allele, indicating that although zygotic Drosophila hunchback is not usually needed to pattern multiple head segments, it is sufficient to do so (in the context of a remaining half dose of bicoid expression). In this comparison, Drosophila hunchback appears to be functionally similar to Nasonia hunchback in the range of segments that it can pattern,although this was not originally obvious from single-mutant analyses.

The roles of genes with overlapping functions, such as orthodenticle and bicoid, have changed during the course of evolution as hunchback has continued to control anterior development. Our finding that hunchback is responsible for controlling more of the anterior development in Nasonia than in Drosophila may indicate that the Hunchback protein has changed its interactions with downstream regulatory genes. Alternatively, the evolution of overlapping gene functions may be sufficient to account for the changing responsibilities of hunchback during the evolution of insect embryos.

We thank Terry Blackman, Loralyn Cozy, Tiana Langan, Megan Lewis and Megan Walker for assistance with embryo collection and stock maintenance. Uyen Tram advised us on fixation techniques and taught us to hand-peel early Nasonia embryos; Jack Werren provided cured wild-type strains and advised us how to cure Nasonia of endosymbiotic bacteria; members of the Small Laboratory advised us on in situ hybridization techniques. We thank Ava Brent, Carol Trent and Ernst Wimmer for providing comments on the manuscript, and Ernst Wimmer for discussing results prior to publication. This work was primarily supported by NSF grant IBN-9808769. Additional support was provided by a WWU pilot project grant to M.A.P. and D.S.L., by a Goldwater Scholarship to S.G., by training grant NIH 5 T32 HD07520 to J.L., and GM064864 to C.D. Sabbatical work of M.A.P. and D.S.L. was partially funded by NSF grants IBN-9982535 to S.S. and IBN-0002958 to C.D., respectively.